Hypothesis-testing abilities of language-impaired children.

Hypothesis-testing abilities were assessed using a modification of the discrimination-learning paradigm employed by Nelson, Kamhi, and Apel (1987) that was designed to minimize the short-term memory demands of the task. Sixteen language-impaired and 16 normal-language children in kindergarten and first and second grades participated in the study. The language-impaired children solved significantly fewer problems than normal-language controls equated on cognitive level, but the two groups used similar hypothesis types to solve the problems. Type of verbal feedback provided during the hypothesis testing task (explicit vs. nonexplicit) did not significantly affect the performance of either group. These results are interpreted as indicating that language-impaired children demonstrate inefficient use of problem-solving strategies that cannot be attributed solely to memory difficulties. Issues surrounding the investigation of language-impaired children's cognitive abilities are discussed.